1. Field of the Invention
This invention relates to a data transmission method and a data processing method of a data processing system comprising a plurality of processing devices.
2. Description of the Prior Art
When a network system, which is connected to a multi-computer system or to a plurality of multi-computer systems with communication lines via gateways, communicates data, one-to-one communication with a destination address assigned to the data or broadcasting with a group address assigned to the data is conventionally performed.
Content code communication indicating that data is assigned a content code indicating the content of the data so that the receiver can decide whether or not to receive the data according to the content code is also available.
These two transmission methods are not used concurrently. A system will use only one transmission method.
In a large scale and wide area system, content code communication using communication lines is expensive except for destination fixed connection.
The communication method using a group address, for example, is indicated in INTAP-S002-02 (V0,R1). Content codes are, for example, indicated in Japanese Patent Publication 59-047905.
In the above prior art, communication using a content code indicating the data content and communication using a destination address or a group address for broadcasting cannot be used concurrently on one transmission line or on one system because they are different from each other in the data message format.
In a large scale and wide area multi-computer system, content code communication broadcasts a data message to all the computers and the receiver decides whether or not to receive the data, causing problems such as a large quantity of data and an increase in cost.
In a multi-computer system, as shown in FIG. 19, a plurality of computers 1901, 1902, 1903, - - - are connected to each other via a transmission line 1910, and transmit or receive data mutually. The computers 1901, 1902, 1903, - - - have external memories 1911, 1912, 1913, - - - for storing data. In the external memories 1911, 1912, 1913, - - - , files A, B, C, D, - - - , where, for example, indexed data is stored, are defined, and multiplexed files exist (for example, file A or file C).
One of the conventional methods which is used for recovering, for example, the file A of the computer 1902 of this multi-computer system, is a method providing that access to the file A is inhibited so that, the file A of the computer 1901 is copied onto the computer 1902, and then access to the file A is permitted after the copy is finished. Since this method inhibits access to the file A, current data of the file A is not generated during recovery, and the processing is simplified.
However, it is inconvenient that the file A cannot be accessed. A file recovery method, which is modified so that the file A of the computer 1901 can be accessed during recovery, is indicated, for example, in Japanese Patent Laid-Open 62-102342.
This recovery method is applied to a sequential file. As shown in FIG. 20(a), the file A of the computer 1901 is copied onto the computer 1902, and current data A.sup.+ generated after the copy starts is added to the computer 1901 as usual but stored in a current data buffer b of the computer 1902. When the file A is copied including up to the data content when the copy starts, the current data A.sup.+ stored in the buffer b is added as shown in FIG. 20(b). By doing this, the file A is recovered including the current data A.sup.+.
In the above recovery method, the recovery processing unit needs a current data buffer for temporarily storing current data during recovery. Since the quantity of current data generated during this recovery period, however, cannot be estimated, an appropriate buffer size cannot be selected. For safety's sake, therefore, it is necessary to install a current data buffer with a capacity larger than the actually required one, causing a problem such as an inefficient facility load. Even if a current data buffer with a large capacity is installed, there is no guarantee that the quantity of current data will not exceed the capacity. If the quantity of current data does, there is a problem imposed that the storage content cannot be recovered. The same may be said with storage content expansion.